Chapter 5. Long-lasting γδ T cell non-responsiveness in patients with Campylobacter jejuni-associated Guillain-Barré syndrome

Transcription

1 Chapter 5 Long-lasting γδ T cell non-responsiveness in patients with Campylobacter jejuni-associated Guillain-Barré syndrome Ildiko Van Rhijn 1, Leonard H. Van den Berg 1, C. Wim Ang 2, and Ton Logtenberg 3,4 Departments of 1 Neurology and 3 Immunology University Medical Center Utrecht, 2 Departments of Neurology and Immunology, Erasmus University Rotterdam, and 4 Crucell, Leiden, The Netherlands 51

2 Abstract Guillain-Barré syndrome (GBS) is an acute, autoimmune neuropathy. Two thirds of the GBS patients report an acute antecedent infectious disease, most frequently Campylobacter jejuni enteritis. Serum antibodies against C. jejuni lipopolysaccharides cross-react with peripheral nerve components, suggesting that molecular mimicry plays a role in the induction of GBS. Here, we aimed to seek further evidence for this hypothesis by studying C. jejuni-reactive T lymphocytes in GBS patients. PBMC of GBS patients and controls were stimulated in vitro with crude sonicates of C. jejuni. In healthy individuals, T lymphocytes expressing γδ but not αβ T cell receptors vigorously proliferated after stimulation with C. jejuni. In contrast, γδ Τ cells of acute GBS patients with antecedent C. jejuni infections completely failed to respond. GBS patients without evidence for antecedent C. jejuni infections and individuals with C. jejuni enteritis without GBS responded like healthy individuals. In some patients, the γδ Τ cell non-responsiveness could last for years after recovery from GBS. Supplementing cell cultures with the cytokines IL-2 or IL-15 resulted in restoration of the γδ Τ cell proliferative response. γδ T cell non-responsiveness in GBS patients reflects a lack of production of cytokines required to activate γδ T cells. T cell nonresponsiveness and ensuing defective autoimmune regulation may be a more general mechanism leading to autoimmune disease. Introduction Guillain-Barré syndrome (GBS) is the leading cause of acute flaccid paralysis in most countries. GBS occurs worldwide and affects patients of all ages. In approximately two-thirds of patients, the neuropathy is preceded by an infectious illness, most commonly a respiratory-tract infection or gastroenteritis. Campylobacter jejuni, a gram-negative rod that is a major cause of bacterial gastroenteritis worldwide, is the most frequent antecedent pathogen An autoimmune etiology of GBS has generally been accepted 141. Antibodies to nerve components have been detected in serum of GBS patients, complement deposits and lymphocytic infiltrates have been observed in sural nerve biopsies and postmortem peripheral nervous system material 1;96;97 and patients have increased levels of circulating IL-2, soluble IL-2 receptors and activated T lymphocytes 4;9;98 Based on the observation that anti C. jejuni lipopolysaccharide antibodies crossreact with gangliosides in peripheral nerve, it has been suggested that molecular mimicry plays a role in the pathogenesis of GBS 94. In this scenario, the immune response to the 52

3 infectious organism would generate antibodies and perhaps T lymphocytes that cross-react with components of peripheral nerve. To gain further insight in the role of T lymphocytes in the pathogenesis of GBS, we analyzed the in vitro response of peripheral blood mononuclear cells from GBS patients with antecedent acute C. jejuni infections and from GBS patients without evidence of recent exposure to this microorganism. As controls, we evaluated the in vitro response of healthy individuals and patients with a history of C. jejuni-induced enteritis. Methods Study subjects An overview of the clinical status of the GBS patients whose blood was collected for this study is presented in Table 1. GBS was diagnosed according to established criteria 142. Acute phase GBS patients were included in the study within one week after the diagnosis of disease. None of the patients received medication at the time blood samples were collected. Experiments with material from a GBS patient and from a healthy control subject were conducted on the same day. None of the healthy control subjects recorded recent symptoms of enteritis. Patients with a history of C. jejuni enteritis, as established by a positive C. jejuni culture from their stool during the acute phase of gastro-enteritis, were also included in the study. The Medical Ethical Committee of the University Medical Center Utrecht approved the study protocol. Serologic studies and criteria for C. jejuni positivity C. jejuni seropositivity was determined in an ELISA as described 91. Acute phase serum of GBS patients was tested for the presence of anti-c. jejuni IgG, IgA, and IgM. Ratios were determined by dividing the OD of the serum by the OD of a negative reference serum representing the cut-off titer. Patients with positive C. jejuni stool culture, patients suffering from diarrheal disease before development of GBS and anti-c. jejuni IgG ratios > 8, and patients with anti-c. jejuni IgM or IgA ratios > 1 were considered as C. jejuni positive. Patients with IgM and IgA ratios < 1 and no history of diarrheal disease were considered to be C. jejuni negative, irrespective of their IgG titer

4 Table 1 Study subjects Clinical status of the GBS and Subject Time after Severity C. jejuni CMV EBV Myco- C. jejuni enteritis patients onset score plasm acute GBS included in this study. The 1/ac acute period between onset of GBS 2 acute acute or C. jejuni enteritis and 4 acute inclusion in our experiments is 5 acute acute depicted as 'Time after onset 7 acute acute disease'. Severity of GBS was 9/ac acute scored at the peak of their 10 acute recovered from GBS illness on a functional 1/ex 2 months disability scale on which years years denotes healthy; 1, minor 13 3 years symptoms; 2, able to walk > years months m without assistance; 3, able to 16 3 years walk > 10 m with support; 4, 17 7 years years bedridden or chairbound; 5, 19 1 month requiring assisted ventilation recovered from C. jejuni enteritis years 4 years /ex 22 7 months 2 years for at least part of the day; 6, dead. Antibodies against CMV, EBV, and mycoplasm were not 23 2 years + determined in C. jejuni 24 1 year years + enteritis patients. T cell cultures and proliferation assays Twenty ml heparinized blood from a GBS patient and from a healthy control subject was collected on the same day and processed in parallel. Blood was diluted 1:1 in PBS and centrifuged over Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) at 900 g during 20 minutes. After two washes in PBS, the cells were counted and resuspended at 10 6 cells/ml in culture medium. Culture medium consisted of RPMI 1640 (Life Technologies, Paisley, Scotland) with L-glutamine, 100 U/ml penicillin, 100 U/ml streptomycin and 5% heat inactivated human pooled serum. An 100 µl aliquot of the PBMC and 100 µl culture medium with antigen or mitogen at double the final concentration were mixed in roundbottom 96-wells microtiter plates and cultured during 12 days prior to cell counting and flow cytometric analysis. For proliferation assays, 1 µci of 3 H labeled thymidine was added to triplicate wells at day 5 of culture. After 18 hours cells were harvested 54

5 onto filter papers and counted in a -counter. Levels of 3 H thymidine incorporation in C. jejunistimulated cultures were calculated by subtracting background 3 H thymidine incorporation from cell cultures containing medium without the crude C. jejuni preparations. The expansion index was calculated by dividing the absolute number of αβ or γδ T cells at day 12 after stimulation by the absolute number at day 0. IL-2 was obtained from Strathmann Biotech (Hannover, Germany) and used at 10 U/ml, and IL-15 was from Pepro Tech (Rocky Hill NJ) and used at 10 ng/ml. Bacterial strains and antigen preparation For stimulation of cells, eight C. jejuni strains were used. Four C. jejuni strains, differing in their serotype as described by Penner 101. were obtained from the ATCC: ATCC 43429: serotype O:1; ATCC 43446: serotype O:19; ATCC 43445: serotype O:18; and CCUG 10938: serotype O:4. In addition, we used four C. jejuni strains isolated from the stool of GBS patients. The serotype of one of these strains could not be determined and the other strains had serotypes O:1, O:2, and O:19. C. jejuni strains were cultured under micro aerobic conditions at 37ºC, on Blood Free Campylobacter Selective Agar Base (Oxoid, Basingstoke, Hampshire, England) supplemented with Cefoperazone (32 mg/litre) and Amphotericine B (5 mg/litre). When growth was sufficient, the bacteria were scraped off the plates and washed twice in PBS. The pellet of approximately 0.5 ml was heated at 100ºC for 40 minutes, followed by sonication for two minutes. The protein concentration of the preparation was assessed by the DC Protein Assay (Bio-Rad, Hercules, CA). The crude C. jejuni antigen preparations were stored at -20ºC until use. Mycobacterium tuberculosis strain H37a was obtained from Difco (Detroit, MI). A vial containing 100 mg of dry material was dissolved in 2 ml water, sonicated, and used at a 1:1000 dilution in T cell cultures. Antibodies and staining procedures Anti-CD3-PECy5, goat anti-mouse-pe, unconjugated anti-fasl, unconjugated anti-tcr ζ chain, unconjugated anti-fasl, and anti-fas-fitc were obtained from Immunotech (Fullerton, CA). Unconjugated anti-tcr Vδ1 (clone A13) was a kind gift of Dr L. Moretta, Genova, unconjugated anti-tcr Vδ2 (clone 4G6), anti-vγ4 (clone 4A11) and anti-vγ9 (clone B3) were generous gifts of Dr G. De Libero, Basel, anti-cd94 was a kind gift of Dr M Lopez Botet, Madrid, and all other antibodies were from Becton Dickinson freshly isolated or cultured PBMC were used per staining. All incubations were on ice. Washes and dilutions of antibodies were carried out with PBS containing 1% BSA. The unconjugated antibodies were used undiluted and removed by washing cells before incubation with goat anti-mouse-fitc (1:50) or goat anti-mouse-pe (1:500). Cells were washed again and incubated with normal mouse serum (1:1000) before other 55

6 antibodies were added. Flow cytometric analysis was performed on a FACSscan (Becton Dickinson, Franklin Lakes, NJ) and for each analysis 5000 ungated events were acquired. In the Figures, the percentage of CD3+ cells is expressed as the percentage of cells within a gate around the living cells. The γδ TCR+ cells and the Vγ and Vδ families are expressed as percentages of the living, CD3+ population. In case the number of γδ T cells detected with the pan-γδ antibody exceeded the sum of the individual Vγ and Vδ families, the presence of γδ T cells expressing a V region not detected with the available monoclonal antibodies was assumed and termed 'undetermined Vγ or Vδ'. Fas, FasL, CD8, HLA-DR, FcRIII, and CD94 positivity is expressed as a percentage of the living, CD3+, γδ TCR+ population. Statistical analyses The mean of triplicate measurements of 3 H thymidine counts and expansion indices was logtransformed prior to statistical evaluation. Values obtained from culture medium control stimulations were subtracted from experimental values. We used a two-tailed Mann-Whithney rank sum test to analyze the differences between healthy control subjects and different patient groups. Results Freshly isolated PBMC were stimulated with crude antigen preparation of eight different C. jejuni strains at a concentration of 3µg protein/ml that, in preliminary dose-response experiments, was determined to be the optimal stimulatory dose for all strains. The eight different C. jejuni strains did not differ in their capacity to stimulate the proliferation of T cells in either patient or healthy control groups (results not shown). Therefore, the experimental values obtained from stimulations with different strains were averaged and are referred to as C. jejuni antigenic stimulation. 3 H- labeled thymidine incorporation was measured at day 5 of the culture, whereas flow cytometric analysis was performed at day 12. PBMC of healthy donors (n = 30) vigorously proliferated upon stimulation with crude C. jejuni preparations (median 30*10 3 cpm, range *10 3 cpm; Figure 1a). In contrast, PBMC of acute GBS patients with a preceding C. jejuni infection (n = 5) did not proliferate upon stimulation with C. jejuni antigen (median 1.2*10 3 cpm, range *10 3 cpm; Figure 1b). The group of acute GBS patients with a preceding C. jejuni infection differed significantly from the group of healthy donors in their proliferative response to C. jejuni antigenic stimulation (P < 0.001). PBMC of GBS patients without evidence for antecedent C. jejuni infections (n = 5), proliferated vigorously in response to C. jejuni stimulation. The magnitude of the response did not 56

7 differ significantly from that of healthy individuals (median 48*10 3 cpm, range *10 3 cpm, Figure 1b). To assess whether the observed non-responsiveness was transient, we collected blood of patient 1 two months after the first experiment. In addition, we analyzed seven patients who had a recorded history of C. jejuni related GBS but were free of symptoms for prolonged periods of time, up till 7 years. Varying responses were observed (n = 8, median 3.9*10 3 cpm, range *10 3 cpm) with two of the eight recovered patients responding similar to the majority of healthy individuals, three patients showing a very low response, and three patients showing absence of a γδ T cell response (Figure 1b). Patients 11 and 17, which were among these low and non-responders, were both free of clinical signs of GBS as long as 7 years prior to these experiments. Patients with a history of gastro-enteritis caused by C. jejuni showed high responses (n = 5, median 52*10 3 cpm, range 32-68*10 3 cpm, Figure 1b ) and differed significantly (P < 0.01) from this group of recovered, C. jejuni positive GBS patients. The uncomplicated gastro-enteritis patients had recovered from their disease between one and three years at the time of inclusion in the experiments A healthy subjects acute GBS C. jejuni pos acute GBS C. jejuni B recovered GBS C. jejuni pos recovered C. jejuni enteritis recovered GBS Figure 1 Proliferative responses of PBMC to C. jejuni antigen Proliferative response to C. jejuni antigens of healthy donors, acute C. jejuni-induced GBS cases, ad acute GBS without preceding C. jejuni infection (a), and patients with a history of Campylobacter enteritis-induced GBS, uncomplicated Campylobacter enteritis, or Campylobacter-unrelated GBS (b). Incorporated 3 H thymidine counts are means of triplicate wells of stimulations with different bacterial strains. 57

8 The phenotype of expanded T cells was determined in subsequent experiments by doubleimmunofluorescent analysis with anti-cd3, anti-tcr αβ, and anti-tcr γδ monoclonal antibodies, and compared to the phenotype of T cells at the onset of cell culture. Flow cytometric analysis demonstrated extensive expansion of CD3+/TCRγδ+ T cells in the cultures of PBMC of all healthy individuals tested. A representative example is shown in Figure 2. The eight C. jejuni strains showed no differences in stimulatory capacity of γδ T cells. CD3 Day 0 CD3 Day 12 1% 35% Pan γδtcr Pan γδtcr Figure 2 γδ T cell expansion after stimulation with C. jejuni antigen Percentages of γδ T cells of a healthy donor at day 0 and day 12 after in vitro stimulation with C. jejuni antigen. Cells are stained with anti-cd3-cy5 and anti-γδ TCR- FITC monoclonal antibodies and prior to flow cytometric analysis. The percentages given are in the living, CD3+ population. In autoimmune and infectious diseases, an aberrant distribution pattern among TCR Vγ/Vδ families and significant expansion or reduction of the entire γδ T cell population has been reported. We therefore analyzed the frequency and Vγ/Vδ utilization of blood γδ T cells of GBS patients and compared the results to those obtained from healthy individuals. In addition, we performed a phenotypic analysis of blood γδ T cells with monoclonal antibodies against molecules that provide information on their activation status, like HLA-DR, anatomical origin, like CD8 143, TCR-mediated signal transduction capacity, like TCRζ chain, CD3, and FcRIII 144, apoptosis related molecules (Fas/FasL) 145, or killer cell inhibitory receptors (CD94) 146. The results show no statistically significant difference between the GBS and healthy control group (not shown). We next determined whether γδ T cells from C. jejuni non-responsive GBS patients were capable of proliferating after stimulation with other antigens. To that end, PBMC of non-responder GBS patients and healthy donors were incubated with a crude preparation of the γδ T cell antigen M. tuberculosis 147. Healthy donors as well as C. jejuni non-responsive GBS patients showed varying γδ T cell responses to M. tuberculosis. The C. jejuni non-responder GBS patient group showed a significantly diminished response (P = 0.044), albeit that some individual patients demonstrated substantial γδ T cell proliferation after stimulation with M. tuberculosis (Figure 3). These experiments suggest that lack of proliferative capacity is not an intrinsic property of the γδ T cells in C. jejuni-related GBS patients. The diminished response in these patients to M. tuberculosis may be explained by shared γδ T cell-stimulatory antigens. 58

9 3 H Thymidine incorporation (log cpm) healthy subj. C. jej non-resp gbs C. jej healthy subj. M. tb non-resp gbs M. tb Figure 3 γδ T cell expansion after stimulation with M. tuberculosis and C. jejuni Responsiveness of healthy subjects and C. jejuni non-responsive GBS patients to M. tuberculosis. The γδ T cell expansion index is plotted on the y- axis. The basis for γδ T cell non-responsiveness was further investigated by adding cytokines to the cell cultures. γδ T cells have been shown to constitutively express the high affinity IL-2 receptor and expand after stimulation with IL ;113 or a combination of antigen and IL-2 or IL ;148;149 PBMC of four non-responding GBS patients were stimulated with C. jejuni, IL-2, IL- 15, or with C. jejuni antigen in combination with these cytokines. After 12 days of culture, expansion indices of γδ T cells were determined. Figure 4 shows that γδ T cells from C. jejuniassociated GBS patients do not respond to stimulation with any of the single compounds, but that stimulation with C. jejuni antigen in combination with IL-2 or with IL-15 results in a strong proliferation, comparable to that observed with PBMC from normal donors. Figure 4 Restoration of the γδ Τ cell proliferative 80 response by cytokines Response of healthy donors (open circles) and Campylobacter non-responsive GBS patients (each patient represented by a different black symbol) to Campylobacter, expansion index IL-2 or IL-15 alone, and Campylobacter antigen in combination with IL-2 or IL-15. The γδ T cell expansion index is plotted on 20 the y-axis. 0 C. jejuni IL-2 C. jejuni + IL-2 IL-15 C. jejuni + IL-15 It has been previously shown that in vitro stimulation of PBMC from M. tuberculosis infected patients with M. tuberculosis antigen induces apoptosis of γδ T lymphocytes within 48 hours 150. A similar mechanism does not appear to be operative in cultures of cells derived from GBS patients stimulated with C. jejuni antigen. Flow cytometric analysis of C. jejuni-stimulated cells at 24 hour 59

10 intervals unveiled that the percentage of γδ T cells within the CD3+ population remained unaltered during the first 7 days in culture (results not shown). Discussion The most salient finding of this study is that γδ T cells in GBS patients with an antecedent acute infection with C. jejuni fail to proliferate after in vitro stimulation with a crude antigen preparation of this micro-organism. This is in striking contrast to the vigorous C. jejuni-induced proliferative response of γδ T cells obtained from healthy individuals, individuals with C. jejunirelated acute enteritis and GBS patients without evidence for an acute antecedent C. jejuni infection. We observed that in some patients, γδ T cell non-responsiveness lasted for prolonged periods of time after clinical recovery. In the most extreme case, the peripheral blood γδ T cells of two patients that had recovered from C. jejuni-related GBS 7 years prior to the time of the experiment, still completely failed to respond. We did not detect quantitative or phenotypic differences in the blood γδ T cell population in GBS patients and healthy controls that could explain the non-responsiveness. PBMC from GBS patients contained normal numbers of circulating γδ T cells that, like γδ T cells in healthy individuals, predominantly expressed TCR encoded by Vγ9/Vδ2 variable genes. Moreover, expression levels of molecules related to γδ T cell activation status and TCR-mediated signal transduction capacity (MHC class II, TCR ζ chain, γδ TCR, CD3, and FcγRIII), apoptosis related molecules (Fas/FasL) or killer cell inhibitory receptors (CD94) were similar among γδ T cells of GBS patients and healthy controls, confirming recently published data 89. Although the γδ T cells did not expand as a result of stimulation with C. jejuni antigen, we observed that they remained present during the culture. The combined data distinguish GBS-related γδ T cell nonresponsiveness from M. tuberculosis related non-responsiveness where stimulation with M. tuberculosis induces Fas/FasL-mediated γδ T cell death and disappearance of γδ T cells from cell cultures 150. Several observations point to a possible mechanism of non-responsiveness. Four out of six C. jejuni non-responder GBS patients also failed to respond to stimulation with preparations of M. tuberculosis, a well-studied antigen inducing γδ T cell proliferation. Thus, non-responsiveness does not appear to be restricted to a C. jejuni specific antigen but extends to a single or a category of γδ T cell stimulatory antigens at least partially shared by M. tuberculosis and C. jejuni bacteria. This hypothesis is further supported by the recent finding that γδ T cells display a diminished 60

11 responsiveness in a non-selected (no discrimination between C. jejuni-related and non-related) group of GBS patients to phosphoantigens, bacterial antigens that potently activate γδ T cells 89. Furthermore, T cells of GBS patients were found to produce little IL-2 upon polyclonal T cell activation and showed impaired responses to IL We noted that γδ T cell responsiveness could be completely restored by simultaneous stimulation of patients PBMC with C. jejuni and the cytokines IL-2 or IL-15. Based on the combined data, we propose that the observed nonresponsiveness in GBS patients is not an inherent property of γδ T cells but rather reflect a deficient cytokine production in the cell cultures, required to drive γδ T cell proliferation. Several reports have shown that γδ T cells can alter antibody responses (reviewed by Born et al. 152 ). Transgenic mice lacking γδ T cells have an exacerbated autoimmune disease phenotype in murine lupus 153, and higher antibody titers against T cell-independent antigens of Borrelia burgdorferi 154 as compared to their littermates with functional γδ T cells. These results suggest that γδ T cells may regulate potentially auto-aggressive immune responses. In GBS, antibodies against C. jejuni have been shown to cross-react with peripheral nerve tissues. The concomitant lack of cytokine production upon C. jejuni stimulation and ensuing γδ non-responsiveness may result in deregulation of autoantibody production and disease. It is tempting to speculate that a similar scenario, based on non-responsiveness of T cells after infection by micro-organisms, unfolds during other autoimmune responses. 61

12 62